Technology

Home
Projects

Leaf Business Consulting Services

Predictive analytics reduces churn

The Challenge

A software company struggled to retain customers through subscription renewals because it could not predict possible customer churn in time. Sales managers and sales executives lacked a central repository of information on past transactions, accounts’ key decision-makers and product-related details and other issues that were necessary for sales expansion.The company reached out a long-standing partner, for help. It had decided to move all its transactional systems

Leaf Business Consulting Services

Adopting ML to Predict Customer Churn

We are dedicated software company team partnered with the client to design and implement a solution that provides 360-degree views of its customers. The solution’s flexible dashboard provides customer performance details at all levels.

Benefits

80% accuracy in customer churn algorithm, 75% to 80% reduction in manual efforts, $12 million in possible cross-sell and upsell opportunities

Our Approach

Using an agile delivery approach, we incorporated machine learning (ML) into the company’s analytics model to elevate its 360-degree view of customers. By applying ML, the client can now proactively take steps to retain customers who are about to discontinue their service and are unlikely to renew their contacts. It can also prioritize high-value customers.

we leveraged Cloudera to host transactional data and Hive to transform the data. By automating the extract, transform and load functions to Cloudera, the client collated data from multiple sources into a central location for easy access and analysis. Preprocessed, customer-centric data provides a 360-degree view of each customer, and that data is fed into the QlikView dashboard. The Apache Spark in-memory processing engine determines customer churn probability and identifies potential cross-sell/upsell opportunities. The flexible solution can be used for approximately 20 product types and 150,000 customers.